Crop-Growth Driven Forward-Modeling of Sentinel-1 Observables Using Machine-Learning

This paper presents an approach to implement a forward model for Sentinel-1 copol and crosspol backscatter and coherence using crop bio-geophysical parameters namely leaf area index, biomass, canopy height, soil moisture and root zone moisture as inputs for the maize. These required input parameters are generated using Decision Support System for Agrotechnology Transfer (DSSAT), one of the state-of-the-art crop growth models. The predicted SAR signal is generated using Support Vector Regression (SVR) over all the maize fields in an agricultural region, Flevoland, Netherlands. The correlation between simulated signal and observed signal is evaluated.